Estimation de densités unimodales
Anne‐Laure Fougères · Canadian Journal of Statistics · 1997
Abstract This paper proposes a new nonparametric unimodal estimator of a unimodal probability density function, in the case where the mode is known. The classical solution to this problem is the maximum‐likelihood estimator under monotonicity constraint, considered by Grenander (1956). Our approach is based on a unimodal rearrangement of the kernel estimator of the density. Asymptotic properties of this estimator are studied, and its small‐sample behaviour is examined through simulations.